Contains 12 datasets for feature selection dimensionality reduction and machine learning. 

The dimension of the selected datasets ranges from 10 to 1000, and the number of instances ranges from 200 to 3000.